gsmopen: added win_iconv

src/mod/endpoints/mod_gsmopen/win_iconv/Makefile

+
+# comma separated list (e.g. "iconv.dll,libiconv.dll")
+DEFAULT_LIBICONV_DLL ?= \"\"
+
+CFLAGS += -pedantic -Wall
+CFLAGS += -DUSE_LIBICONV_DLL
+CFLAGS += -DDEFAULT_LIBICONV_DLL=$(DEFAULT_LIBICONV_DLL)
+
+all: iconv.dll libiconv.a win_iconv.exe
+
+dist: test win_iconv.zip
+
+iconv.dll: win_iconv.c
+	gcc $(CFLAGS) -c win_iconv.c -DMAKE_DLL
+	dllwrap --dllname iconv.dll --def iconv.def win_iconv.o $(SPECS_FLAGS)
+	strip iconv.dll
+
+libiconv.a: win_iconv.c
+	gcc $(CFLAGS) -c win_iconv.c
+	ar rcs libiconv.a win_iconv.o
+	ranlib libiconv.a
+
+win_iconv.exe: win_iconv.c
+	gcc $(CFLAGS) -s -o win_iconv.exe win_iconv.c -DMAKE_EXE
+
+libmlang.a: mlang.def
+	dlltool --kill-at --input-def mlang.def --output-lib libmlang.a
+
+test:
+	gcc $(CFLAGS) -s -o win_iconv_test.exe win_iconv_test.c
+	./win_iconv_test.exe
+
+win_iconv.zip: msvcrt msvcr70 msvcr71
+	rm -rf win_iconv
+	svn export . win_iconv
+	cp msvcrt/iconv.dll msvcrt/win_iconv.exe win_iconv/
+	mkdir win_iconv/msvcr70
+	cp msvcr70/iconv.dll win_iconv/msvcr70/
+	mkdir win_iconv/msvcr71
+	cp msvcr71/iconv.dll win_iconv/msvcr71/
+	zip -r win_iconv.zip win_iconv
+
+msvcrt:
+	svn export . msvcrt; \
+	cd msvcrt; \
+	$(MAKE);
+
+msvcr70:
+	svn export . msvcr70; \
+	cd msvcr70; \
+	gcc -dumpspecs | sed s/-lmsvcrt/-lmsvcr70/ > specs; \
+	$(MAKE) "SPECS_FLAGS=-specs=$$PWD/specs";
+
+msvcr71:
+	svn export . msvcr71; \
+	cd msvcr71; \
+	gcc -dumpspecs | sed s/-lmsvcrt/-lmsvcr71/ > specs; \
+	$(MAKE) "SPECS_FLAGS=-specs=$$PWD/specs";
+
+clean:
+	rm -f win_iconv.exe
+	rm -f win_iconv.o
+	rm -f iconv.dll
+	rm -f libiconv.a
+	rm -f win_iconv_test.exe
+	rm -f libmlang.a
+	rm -rf win_iconv
+	rm -rf win_iconv.zip
+	rm -rf msvcrt
+	rm -rf msvcr70
+	rm -rf msvcr71
+

src/mod/endpoints/mod_gsmopen/win_iconv/iconv.def

+LIBRARY ICONV.DLL
+EXPORTS
+  iconv
+  iconv_open
+  iconv_close
+  iconvctl
+  libiconv=iconv
+  libiconv_open=iconv_open
+  libiconv_close=iconv_close
+  libiconvctl=iconvctl
+;; libiconv-1.11.dll
+;; TODO for binary compatibility
+;  _libiconv_version @1
+;  aliases2_lookup @2
+;  aliases_lookup @3
+;  iconv_canonicalize @4
+;  libiconv @5
+;  libiconv_close @6
+;  libiconv_open @7
+;  libiconv_relocate @8
+;  libiconv_set_relocation_prefix @9
+;  libiconvctl @10
+;  libiconvlist @11
+;  locale_charset @12

src/mod/endpoints/mod_gsmopen/win_iconv/iconv.h

+#include <stddef.h>
+typedef void* iconv_t;
+iconv_t iconv_open(const char *tocode, const char *fromcode);
+int iconv_close(iconv_t cd);
+size_t iconv(iconv_t cd, const char **inbuf, size_t *inbytesleft, char **outbuf, size_t *outbytesleft);

src/mod/endpoints/mod_gsmopen/win_iconv/mlang.def

+LIBRARY MLANG.DLL
+EXPORTS
+  ConvertINetMultiByteToUnicode@24
+  ;; ConvertINetReset (not documented)
+  ConvertINetString@28
+  ConvertINetUnicodeToMultiByte@24
+  IsConvertINetStringAvailable@8
+  LcidToRfc1766A@12
+  LcidToRfc1766W@12
+  Rfc1766ToLcidA@8
+  Rfc1766ToLcidW@8

src/mod/endpoints/mod_gsmopen/win_iconv/mlang.h

+HRESULT WINAPI ConvertINetString(
+    LPDWORD lpdwMode,
+    DWORD dwSrcEncoding,
+    DWORD dwDstEncoding,
+    LPCSTR lpSrcStr,
+    LPINT lpnSrcSize,
+    LPBYTE lpDstStr,
+    LPINT lpnDstSize
+);
+
+HRESULT WINAPI ConvertINetMultiByteToUnicode(
+    LPDWORD lpdwMode,
+    DWORD dwSrcEncoding,
+    LPCSTR lpSrcStr,
+    LPINT lpnMultiCharCount,
+    LPWSTR lpDstStr,
+    LPINT lpnWideCharCount
+);
+
+HRESULT WINAPI ConvertINetUnicodeToMultiByte(
+    LPDWORD lpdwMode,
+    DWORD dwEncoding,
+    LPCWSTR lpSrcStr,
+    LPINT lpnWideCharCount,
+    LPSTR lpDstStr,
+    LPINT lpnMultiCharCount
+);
+
+HRESULT WINAPI IsConvertINetStringAvailable(
+    DWORD dwSrcEncoding,
+    DWORD dwDstEncoding
+);
+
+HRESULT WINAPI LcidToRfc1766A(
+    LCID Locale,
+    LPSTR pszRfc1766,
+    int nChar
+);
+
+HRESULT WINAPI LcidToRfc1766W(
+    LCID Locale,
+    LPWSTR pszRfc1766,
+    int nChar
+);
+
+HRESULT WINAPI Rfc1766ToLcidA(
+    LCID *pLocale,
+    LPSTR pszRfc1766
+);
+
+HRESULT WINAPI Rfc1766ToLcidW(
+    LCID *pLocale,
+    LPWSTR pszRfc1766
+);

src/mod/endpoints/mod_gsmopen/win_iconv/readme.txt

+win_iconv is a iconv library using Win32 API to conversion.
+win_iconv is placed in the public domain.
+Yukihiro Nakadaira <yukihiro.nakadaira@gmail.com>

src/mod/endpoints/mod_gsmopen/win_iconv/win_iconv_test.c

+
+#include "win_iconv.c"
+#include <stdio.h>
+
+const char *
+tohex(const char *str, int size)
+{
+    static char buf[BUFSIZ];
+    char *pbuf = buf;
+    int i;
+    buf[0] = 0;
+    for (i = 0; i < size; ++i)
+        pbuf += sprintf(pbuf, "%02X", str[i] & 0xFF);
+    return buf;
+}
+
+const char *
+errstr(int errcode)
+{
+    static char buf[BUFSIZ];
+    switch (errcode)
+    {
+    case 0: return "NOERROR";
+    case EINVAL: return "EINVAL";
+    case EILSEQ: return "EILSEQ";
+    case E2BIG: return "E2BIG";
+    }
+    sprintf(buf, "%d\n", errcode);
+    return buf;
+}
+
+#ifdef USE_LIBICONV_DLL
+int use_dll;
+
+int
+setdll(const char *dllpath)
+{
+    char buf[BUFSIZ];
+    rec_iconv_t cd;
+
+    sprintf(buf, "WINICONV_LIBICONV_DLL=%s", dllpath);
+    putenv(buf);
+    if (libiconv_iconv_open(&cd, "ascii", "ascii"))
+    {
+        FreeLibrary(cd.hlibiconv);
+        use_dll = TRUE;
+        return TRUE;
+    }
+    use_dll = FALSE;
+    return FALSE;
+}
+#endif
+
+/*
+ * We can test the codepage that is installed in the system.
+ */
+int
+check_enc(const char *encname, int codepage)
+{
+    iconv_t cd;
+    int cp;
+    cd = iconv_open("utf-8", encname);
+    if (cd == (iconv_t)(-1))
+    {
+        printf("%s(%d) IS NOT SUPPORTED: SKIP THE TEST\n", encname, codepage);
+        return FALSE;
+    }
+    cp = ((rec_iconv_t *)cd)->from.codepage;
+    if (cp != codepage)
+    {
+        printf("%s(%d) ALIAS IS MAPPED TO DIFFERENT CODEPAGE (%d)\n", encname, codepage, cp);
+        exit(1);
+    }
+    iconv_close(cd);
+    return TRUE;
+}
+
+int use_dll;
+
+void
+test(const char *from, const char *fromstr, int fromsize, const char *to, const char *tostr, int tosize, int errcode, int bufsize, int line)
+{
+    char outbuf[BUFSIZ];
+    const char *pin;
+    char *pout;
+    size_t inbytesleft;
+    size_t outbytesleft;
+    iconv_t cd;
+    size_t r;
+    char dllpath[_MAX_PATH];
+
+    cd = iconv_open(to, from);
+    if (cd == (iconv_t)(-1))
+    {
+        printf("%s -> %s: NG: INVALID ENCODING NAME: line=%d\n", from, to, line);
+        exit(1);
+    }
+
+#ifdef USE_LIBICONV_DLL
+    if (((rec_iconv_t *)cd)->hlibiconv != NULL)
+        GetModuleFileName(((rec_iconv_t *)cd)->hlibiconv, dllpath, sizeof(dllpath));
+
+    if (use_dll && ((rec_iconv_t *)cd)->hlibiconv == NULL)
+    {
+        printf("%s: %s -> %s: NG: FAILED TO USE DLL: line=%d\n", dllpath, from, to, line);
+        exit(1);
+    }
+    else if (!use_dll && ((rec_iconv_t *)cd)->hlibiconv != NULL)
+    {
+        printf("%s: %s -> %s: NG: DLL IS LOADED UNEXPECTEDLY: line=%d\n", dllpath, from, to, line);
+        exit(1);
+    }
+#endif
+
+    errno = 0;
+
+    pin = fromstr;
+    pout = outbuf;
+    inbytesleft = fromsize;
+    outbytesleft = bufsize;
+    r = iconv(cd, &pin, &inbytesleft, &pout, &outbytesleft);
+    if (r != (size_t)(-1))
+        r = iconv(cd, NULL, NULL, &pout, &outbytesleft);
+    *pout = 0;
+
+#ifdef USE_LIBICONV_DLL
+    if (use_dll)
+        printf("%s: ", dllpath);
+#endif
+    printf("%s(%s) -> ", from, tohex(fromstr, fromsize));
+    printf("%s(%s%s%s): ", to, tohex(tostr, tosize),
+            errcode == 0 ? "" : ":",
+            errcode == 0 ? "" : errstr(errcode));
+    if (strcmp(outbuf, tostr) == 0 && errno == errcode)
+        printf("OK\n");
+    else
+    {
+        printf("RESULT(%s:%s): ", tohex(outbuf, sizeof(outbuf) - outbytesleft),
+                errstr(errno));
+        printf("NG: line=%d\n", line);
+        exit(1);
+    }
+}
+
+#define STATIC_STRLEN(arr) (sizeof(arr) - 1)
+
+#define success(from, fromstr, to, tostr) test(from, fromstr, STATIC_STRLEN(fromstr), to, tostr, STATIC_STRLEN(tostr), 0, BUFSIZ, __LINE__)
+#define einval(from, fromstr, to, tostr) test(from, fromstr, STATIC_STRLEN(fromstr), to, tostr, STATIC_STRLEN(tostr), EINVAL, BUFSIZ, __LINE__)
+#define eilseq(from, fromstr, to, tostr) test(from, fromstr, STATIC_STRLEN(fromstr), to, tostr, STATIC_STRLEN(tostr), EILSEQ, BUFSIZ, __LINE__)
+#define e2big(from, fromstr, to, tostr, bufsize) test(from, fromstr, STATIC_STRLEN(fromstr), to, tostr, STATIC_STRLEN(tostr), E2BIG, bufsize, __LINE__)
+
+int
+main(int argc, char **argv)
+{
+#ifdef USE_LIBICONV_DLL
+    /* test use of dll if $DEFAULT_LIBICONV_DLL was defined. */
+    if (setdll(""))
+    {
+        success("ascii", "ABC", "ascii", "ABC");
+        success("ascii", "ABC", "utf-16be", "\x00\x41\x00\x42\x00\x43");
+    }
+    else
+    {
+        printf("\nDLL TEST IS SKIPPED\n\n");
+    }
+
+    setdll("none");
+#endif
+
+    if (check_enc("ascii", 20127))
+    {
+        success("ascii", "ABC", "ascii", "ABC");
+        /* MSB is dropped.  Hmm... */
+        success("ascii", "\x80\xFF", "ascii", "\x00\x7F");
+    }
+
+    /* unicode (CP1200 CP1201 CP12000 CP12001 CP65001) */
+    if (check_enc("utf-8", 65001)
+            && check_enc("utf-16be", 1201) && check_enc("utf-16le", 1200)
+            && check_enc("utf-32be", 12001) && check_enc("utf-32le", 12000)
+            )
+    {
+        /* Test the BOM behavior
+         * 1. Remove the BOM when "fromcode" is utf-16 or utf-32.
+         * 2. Add the BOM when "tocode" is utf-16 or utf-32.  */
+        success("utf-16", "\xFE\xFF\x01\x02", "utf-16be", "\x01\x02");
+        success("utf-16", "\xFF\xFE\x02\x01", "utf-16be", "\x01\x02");
+        success("utf-32", "\x00\x00\xFE\xFF\x00\x00\x01\x02", "utf-32be", "\x00\x00\x01\x02");
+        success("utf-32", "\xFF\xFE\x00\x00\x02\x01\x00\x00", "utf-32be", "\x00\x00\x01\x02");
+        success("utf-16", "\xFE\xFF\x00\x01", "utf-8", "\x01");
+#ifndef GLIB_COMPILATION
+        success("utf-8", "\x01", "utf-16", "\xFE\xFF\x00\x01");
+        success("utf-8", "\x01", "utf-32", "\x00\x00\xFE\xFF\x00\x00\x00\x01");
+#else
+        success("utf-8", "\x01", "utf-16", "\xFF\xFE\x01\x00");
+        success("utf-8", "\x01", "utf-32", "\xFF\xFE\x00\x00\x01\x00\x00\x00");
+#endif
+
+        success("utf-16be", "\xFE\xFF\x01\x02", "utf-16be", "\xFE\xFF\x01\x02");
+        success("utf-16le", "\xFF\xFE\x02\x01", "utf-16be", "\xFE\xFF\x01\x02");
+        success("utf-32be", "\x00\x00\xFE\xFF\x00\x00\x01\x02", "utf-32be", "\x00\x00\xFE\xFF\x00\x00\x01\x02");
+        success("utf-32le", "\xFF\xFE\x00\x00\x02\x01\x00\x00", "utf-32be", "\x00\x00\xFE\xFF\x00\x00\x01\x02");
+        success("utf-16be", "\xFE\xFF\x00\x01", "utf-8", "\xEF\xBB\xBF\x01");
+        success("utf-8", "\xEF\xBB\xBF\x01", "utf-8", "\xEF\xBB\xBF\x01");
+
+        success("utf-16be", "\x01\x02", "utf-16le", "\x02\x01");
+        success("utf-16le", "\x02\x01", "utf-16be", "\x01\x02");
+        success("utf-16be", "\xFE\xFF", "utf-16le", "\xFF\xFE");
+        success("utf-16le", "\xFF\xFE", "utf-16be", "\xFE\xFF");
+        success("utf-32be", "\x00\x00\x03\x04", "utf-32le", "\x04\x03\x00\x00");
+        success("utf-32le", "\x04\x03\x00\x00", "utf-32be", "\x00\x00\x03\x04");
+        success("utf-32be", "\x00\x00\xFF\xFF", "utf-16be", "\xFF\xFF");
+        success("utf-16be", "\xFF\xFF", "utf-32be", "\x00\x00\xFF\xFF");
+        success("utf-32be", "\x00\x01\x00\x00", "utf-16be", "\xD8\x00\xDC\x00");
+        success("utf-16be", "\xD8\x00\xDC\x00", "utf-32be", "\x00\x01\x00\x00");
+        success("utf-32be", "\x00\x10\xFF\xFF", "utf-16be", "\xDB\xFF\xDF\xFF");
+        success("utf-16be", "\xDB\xFF\xDF\xFF", "utf-32be", "\x00\x10\xFF\xFF");
+        eilseq("utf-32be", "\x00\x11\x00\x00", "utf-16be", "");
+        eilseq("utf-16be", "\xDB\xFF\xE0\x00", "utf-32be", "");
+        success("utf-8", "\xE3\x81\x82", "utf-16be", "\x30\x42");
+        einval("utf-8", "\xE3", "utf-16be", "");
+    }
+
+    /* Japanese (CP932 CP20932 CP50220 CP50221 CP50222 CP51932) */
+    if (check_enc("cp932", 932)
+            && check_enc("cp20932", 20932) && check_enc("euc-jp", 51932)
+            && check_enc("cp50220", 50220) && check_enc("cp50221", 50221)
+            && check_enc("cp50222", 50222) && check_enc("iso-2022-jp", 50221))
+    {
+        /* Test the compatibility for each other Japanese codepage.
+         * And validate the escape sequence handling for iso-2022-jp. */
+        success("utf-16be", "\xFF\x5E", "cp932", "\x81\x60");
+        success("utf-16be", "\x30\x1C", "cp932", "\x81\x60");
+        success("utf-16be", "\xFF\x5E", "cp932//nocompat", "\x81\x60");
+        eilseq("utf-16be", "\x30\x1C", "cp932//nocompat", "");
+        success("euc-jp", "\xA4\xA2", "utf-16be", "\x30\x42");
+        einval("euc-jp", "\xA4\xA2\xA4", "utf-16be", "\x30\x42");
+        eilseq("euc-jp", "\xA4\xA2\xFF\xFF", "utf-16be", "\x30\x42");
+        success("cp932", "\x81\x60", "iso-2022-jp", "\x1B\x24\x42\x21\x41\x1B\x28\x42");
+        success("UTF-16BE", "\xFF\x5E", "iso-2022-jp", "\x1B\x24\x42\x21\x41\x1B\x28\x42");
+        eilseq("UTF-16BE", "\x30\x1C", "iso-2022-jp//nocompat", "");
+        success("UTF-16BE", "\x30\x42\x30\x44", "iso-2022-jp", "\x1B\x24\x42\x24\x22\x24\x24\x1B\x28\x42");
+        success("iso-2022-jp", "\x1B\x24\x42\x21\x41\x1B\x28\x42", "UTF-16BE", "\xFF\x5E");
+    }
+
+    /*
+     * test for //translit
+     * U+FF41 (FULLWIDTH LATIN SMALL LETTER A) <-> U+0062 (LATIN SMALL LETTER A)
+     */
+    eilseq("UTF-16BE", "\xFF\x41", "iso-8859-1", "");
+    success("UTF-16BE", "\xFF\x41", "iso-8859-1//translit", "a");
+
+    /*
+     * TODO:
+     * Test for state after iconv() failed.
+     * Ensure iconv() error is safe and continuable.
+     */
+
+    return 0;
+}
+